Homing Trajectories and Initial Orientation in a Neotropical

Pašukonis et al. Frontiers in Zoology 2014, 11:29 http://www.frontiersinzoology.com/content/11/1/29

RESEARCH Open Access Homing trajectories and initial orientation in a Neotropical territorial frog, Allobates femoralis (Dendrobatidae) Andrius Pašukonis1*, Matthias-Claudio Loretto1, Lukas Landler2, Max Ringler3,4 and Walter Hödl3

Abstract Introduction: The ability to relocate home or breeding sites after experimental removal has been observed in several amphibians and the sensory basis of this behavior has been studied in some temperate-region species. However, the actual return trajectories have rarely been quantified in these studies and it remains unknown how different cues guide the homing behavior. Dendrobatidae (dart-poison frogs) exhibit some of the most complex spatial behaviors among amphibians, such as territoriality and tadpole transport. Recent data showed that Allobates femoralis, a frog with paternal tadpole transport, successfully returns to the home territories after experimental translocations of up to 400 m. In the present study, we used harmonic direction finding to obtain homing trajectories. Additionally, we quantified the initial orientation of individuals, translocated 10 m to 105 m, in an arena assay. Results: Tracking experiments revealed that homing trajectories are characterized by long periods of immobility (up to several days) and short periods (several hours) of rapid movement, closely fitting a straight line towards the home territory. In the arena assay, the frogs showed significant homeward orientation for translocation distances of 35 m to 70 m but not for longer and shorter distances. Conclusions: Our results describe a very accurate homing behavior in male A. femoralis. The straightness of trajectories and initial homeward orientation suggest integration of learned landmarks providing a map position for translocated individuals. Future research should focus on the role of learning in homing behavior and the exact nature of cues being used. Keywords: Homing, Orientation, Telemetry, Dendrobatidae, Allobates femoralis

Introduction observed in many amphibians (e.g., Anaxyrus terrestris The ability to quantify individual movement patterns (Bufonidae) [7], Lithobates pipiens (Ranidae) [8], Pseuda- using modern telemetry techniques has been crucial in cris regilla (Hylidae) [9], Allobates femoralis (Dendrobati- understanding animal orientation. Homing ability and tra- dae) [10], Taricha rivularis (Salamandridae) [11]). Several jectories following translocations from home or breeding sensory modalities have been implicated in this ability, sites have been widely used to study the orientation mech- most commonly olfaction [12-14], magneto-reception anisms involved (e.g., homing pigeons [1], desert ants [2]). [15,16], and a celestial compass [5,17]. Often, integration Directed, relatively long distance movements, such as and compensation of different modalities depending on mass spring migration to breeding ponds in temperate- cue availability have been suggested (reviewed in [6,18]). region amphibians, has historically drawn much atten- Even though the importance of certain sensory modalities tion to amphibian orientation ([3,4], reviewed in [5,6]). has been revealed for some temperate-region amphibians, A tendency to move towards and to relocate home or the exact homing trajectories have not been quantified breeding sites after experimental removal has been (but see [3,4,18,19]). Thus it remains unknown how exactly these cues guide the behavior. * Correspondence: [email protected] Traditionally, orientation mechanisms in amphibians 1Department of Cognitive Biology, University of Vienna, Althanstrasse 14, 1090 Vienna, Austria have been studied by observing initial orientation over a Full list of author information is available at the end of the article few meters in arena setups or over longer distances

© 2014 Pašukonis et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-2-15l805ox7ya4a9 Pašukonis et al. Frontiers in Zoology 2014, 11:29 Page 2 of 9 http://www.frontiersinzoology.com/content/11/1/29

using drift fences and pit-fall traps. In an attempt to individuals translocated for 50 m, 12 successfully re- track individuals, several authors have attached trailing turned to their home territories. Three individuals did devices to the backs of larger anurans, i.e. a thread bob- not move from their release sites for three to four days, bin, which releases a thread as the animal moves after which we removed their tags and returned them to [8,13,20,21]. However, this tracking method hinders ani- their home territory to avoid any long-term effect on in- mal movements, often results in injuries or a complete dividuals’ health and behavior. Two individuals lost the inhibition of homing [8,22], and is limited to large spe- reflector antenna and could not be located en route but cies. The miniaturization of radio transmitters has were recovered back in their home territories. Ten indi- strongly expanded the range of species suitable for tel- viduals were successfully tracked en route. On average emetry, including many amphibians [23,24]. In parallel, the trajectory straightness coefficient was very high harmonic radar [25] and harmonic direction-finding (MeanSC = 0.93, SD = 0.075) and the inferred homing techniques [26] were developed, allowing to track even trajectories were closely fitting a straight homewards line some of the smallest amphibians. Unfortunately, limited (Figure 1). Translocated individuals showed a significant detection range and problematic external attachment on initial homeward orientation (Mean = 344.5°, 95% CI = amphibians have resulted in predominantly methodo- 304.7° – 24.3°, Rayleigh-test p = 0.01, n = 8). logical studies (e.g., [27-29], but see [30,31]). Telemetry Total return time ranged from 4 h to 102 h (Mean = techniques have rarely been used to gain new insights 37.82 h, SD = 29.49). During their diurnal activity into amphibian orientation. period, individuals spent significantly more time immo- Orientation has mostly been studied in nocturnal bile than moving (medianmoving = 4.88 h, Q1–Q3moving = temperate-region anurans, especially bufonids, but it is 4–8.06 h, medianstationary = 14.88 h, Q1–Q3stationary = in tropical amphibians that we find some of the most 4.38–23.13 h, Z = −1.97, Wilcoxon p = 0.05, n = 10). As complex spatial behaviors. Dart-poison frogs (Dendroba- seen from the quartile values, there was relatively little tidae) are a group of diurnal Neotropical frogs character- variation in the active movement time necessary for ized by territoriality and parental care, which includes frogs to return to their home territory but a much tadpole transport [32]. Despite the fact that dendrobatid greater variation in the time they spent immobile. parental care has been investigated in great detail [32-34], very little is known about the related spatial be- Arena assay haviors. Allobates femoralis is a small dendrobatid frog A second order circular tests for unimodal distribution common throughout the Amazon basin and the Guiana revealed a significant homeward orientation for the Shield [35]. At the onset of the rainy season, males es- tablish territories, which are vocally advertised and Home territory defended for up to several months [36,37]. Courtship, mating and oviposition take place in the leaf-litter within the male’s territory. Tadpoles are later transported to widely dispersed aquatic deposition sites, such as tem- porary pools in flooded areas or holes in fallen trees, as far as 185 m away from the territory [38]. Recent data have shown that experimentally translocated territorial males can return from up to 400 m with increased hom- 270° 90° ing success rates under 200 m [10]. 50 25 0 25 50 m In the present study we further describe the homing behavior of male A. femoralis with a specific focus on the actual movement patterns during return to their home territory. We used telemetry with miniature passive transponders to quantify the homing trajectories of territorial males, displaced over 50 m. Additionally, we translocated males from their home territories into a circular arena within their natural habitat and quantified 180° the initial movements after release. Figure 1 Homing trajectories circular. Circular plot showing the homing trajectories of ten territorial males that were translocated Results over 50 m in the telemetry trials. For better visualization, all Telemetry experiment trajectories were normalized to a single starting point and axis. Each Fifteen out of 17 male frogs equipped with transponders line represents a different individual. Points mark en route locations connected by linear interpolation. continued to show territorial behavior. Out of 15 Pašukonis et al. Frontiers in Zoology 2014, 11:29 Page 3 of 9 http://www.frontiersinzoology.com/content/11/1/29

individuals from the mid-range (n = 13, Mean vector = significant orientation for translocation distances of 345.1°, 95% CI = 275.7° – 49.1°, Hotelling’s F = 5.49, p = 35 m to 70 m but not for longer and shorter distances. 0.02) but not for the close-range (n = 14, Hotelling’sp> The most intriguing finding is the accuracy of the ob- 0.1) and far-range samples (n = 11, Hotelling’s p > 0.1) served homing trajectories. The straightness of trajector- (Figure 2). Confirming the between-group difference, ies over similar distances is more comparable with path there was a significant difference in distribution between integration based homing in desert ants (e.g., [2]) or dir- the mid-range and the close-range (Hotelling’s two sam- ect guidance in social insects (e.g., [39]) rather than any ple F = 4.06, p = 0.03) as well as mid-range and far-range homing trajectories described in amphibians [8,40]. samples (Hotelling’s two sample F = 4.64, p = 0.02) but However, experimental translocation from territories ex- not between the close and far-range samples (Hotelling’s cludes the possibility of path integration in our study. p > 0.1). Overall, we could not unambiguously confirm a Direct guidance (also termed beaconing) requires direct successful homing after the trial for six individuals (three sensory contact to the goal. We can exclude direct visual for the close-range and three for the-far range). contact with the goal in the dense rainforest. The role of olfaction in orientation has been demonstrated for several anuran species, but in these cases animals were Discussion orienting in an open landscape to a large goal such as a Our tracking experiment revealed that homing trajector- breeding pond [12-14,20]. It is difficult to imagine how ies of translocated male A. femoralis are characterized by olfactory guidance would explain such an accurate hom- rapid movements, closely fitting a straight line towards ing to small terrestrial territories in a forest understory the home territory, with occasional long periods of im- with hardly any stable winds. Alternatively, it has been mobility lasting up to several days (Figure 3). We did suggested that some animals could form ‘olfactory maps’, not observe any search patterns and the initial move- which allow them to navigate an area, but the empirical ments over several meters after release were already ori- evidence is equivocal [41]. Magnetic compass and mag- ented homewards. Detailed analysis of the very first netic maps have been implicated in long distance hom- movements within one meter in an arena setup revealed ing of several species [42], including urodele amphibians

(b) 35 - 70 m Home direction

270° 90° 1 0.5 0.5 1

p = 0.02

(a) < 35 m p = 0.03 180° p = 0.02 (c) 70 - 105 m

n. s. n. s.

270° n. s. 90° 1 0.5 0.5 1 1 0.5 0.5 1

180° 180° Figure 2 Arena orientation. Circular vector plots representing initial orientation over one meter of territorial individuals translocated from close- range, < 35 m (a), mid-range, 35 – 70 m (b), or far-range, 70 – 105 m (c). Each vector represents an individual. Vector length represents the path straightness in the arena. Vector direction represents the bearing from the arena center to the crossing point of a one-meter radius circle. Reported significance levels were obtained by second order Hotelling’s circlular test for unimodal distribution and by a Hotelling’s two-sample test for the comparisons between distributions. Pašukonis et al. Frontiers in Zoology 2014, 11:29 Page 4 of 9 http://www.frontiersinzoology.com/content/11/1/29

Figure 3 Trajectory map. Map of the study area showing homing trajectories and temporal distribution of ten territorial males, each individual represented by a different color. Square frog symbols indicate the home territories; x-symbols show release points; circles show en route locations connected with linear interpolation lines. The size of each circle is proportional to the time spent at each location. Contour lines (1 m) are in light gray, creeks and Arataye River in dark gray.

[15,16], but the shallow magnetic gradients of the earth’s homing behavior. More generally, many dendrobatid magnetic field could not account for the observed hom- frogs are likely to rely on spatial learning for finding ing precision at these relatively short distances [43]. small and widely dispersed aquatic tadpole deposition Acoustic orientation has been well documented for male sites in complex environment but research in this area is A. femoralis in a territorial defense context [44-46] and lacking. females most likely use males’ advertisement calls as The long immobility periods shown by most individ- beacons to find territory owners over several tens of me- uals before moving could be related to the time neces- ters [47]. We speculate that translocated individuals sary to accumulate or to perceive temporally varying could use other territorial males as acoustic beacons or orientation cues, such as conspecific vocalizations. We even integrate their calling positions and identities into a did not observe any clear relation between frogs’ activity full acoustic map of the area. To the best of our know- during homing and the time of the day or the amount of ledge, the potential existence of such orientation mech- precipitation. We do not think that catching or tag at- anism has not yet been investigated. Broad acoustic tachment induced stress caused these latencies. Our ex- gradients such as provided by a nearby river could be perience from a long-term recapture study of A. used in addition. The function of broad acoustic gradi- femoralis has shown that most males resume their previ- ents in homing has been often discussed in anuran ous activities such as calling or courtship after being orientation, usually in the context of a breeding chorus caught or even toe-clipped ([49] and personal obs. by A. (e.g., [7,14,48]), but there is little evidence of acoustics Pašukonis). Further, the average return time from 50 m being a primary cue in anuran homing. for non-tagged animals (2 days, [10]) is similar to the Other types of landmark-based orientation are also one observed for tagged individuals in this study (38 h). possible. However, the lack of any observable search pat- Of course, it is possible that the removal from a familiar terns and strong initial orientation suggest, that if spatial location results in stress induced passiveness and immo- learning is involved, landmarks may be integrated on an bility. However, we regularly observed translocated frogs internal map and not simply used as intermediate bea- waiting for hours on exposed, slightly elevated structures cons. Translocation distances of 50 m are well inside the at the release site, which is an indication of an alert state potential range explored by most individuals during tad- rather than passiveness. pole transport [38] or juvenile dispersal (M. Ringler, R. Initial movements of frogs translocated into the arena Mangione, A. Pašukonis, E. Ringler, unpublished data) for distances comparable to the tracking experiment and spatial information could be acquired during these (35–70 m) support the assumption that A. femoralis per- movements. Homing success in A. femoralis is very high ceives some cues directly at the release site. However, up to 200 m [10], which corresponds to the maximum the overall results are equivocal, as the individuals with distances covered by tadpole transporting adults, and territories closer than 35 m and further than 70 m to the further suggests the role of familiarity with the area in arena did not show significant homeward orientation. A Pašukonis et al. Frontiers in Zoology 2014, 11:29 Page 5 of 9 http://www.frontiersinzoology.com/content/11/1/29

strong motivation to escape from the arena might over- landmarks providing a map position for translocated in- ride the immediate homing motivation. More specific- dividuals. We suggest that spatial knowledge acquired ally, frogs displaced over very short distances might have during juvenile dispersal or tadpole transport mediates had less motivation to orient towards their territory be- this behavior but the actual cues being used remain un- fore leaving the arena. On the other hand, individuals known. Future research should focus on the importance coming from further than 70 m might have not been of familiarity with an area for successful homing and the able to perceive the orientation cue within the given trial nature of sensory cues being used. time. In the arena, most individuals moved within the first 90 min and the remaining few were excluded from Materials and methods the trial, while after the release at natural sites in the Study animals and area tracking experiment, some frogs remained stationary for Allobates femoralis is a small (snout-urostyle length ap- more than 24 h before initiating their movements. This proximately 25–30 mm) territorial dendrobatid frog. indicates, that if the orientation cues are only temporally Playback of an advertisement call of a simulated intruder available or need to be accumulated for correct orienta- reliably elicits antiphonal calling or direct phonotactic tion, the arena setup was not suitable to test for this approach by the resident male [44]. Individual frogs can ability. It is important to note, that even the males that be identified and recognized by their unique ventral col- were displaced to the arena from longer distance oration patterns [49]. returned to their home territories. Additional tracking of The study was carried out within one reproductive individuals after testing them in the arena would be ne- season of A. femoralis between 19 January 2013 and 30 cessary to understand how the direction choice in the March 2013. Frogs were sampled from a single popula- arena relates to the overall homing motivation and tion in an area of approximately 3 ha near the field camp trajectories. ‘Saut Pararé’ (4°02’ N, 52°41’ W, WGS84) in the nature To the best of our knowledge, this is the first study reserve ‘Les Nouragues’, French Guiana. The study area quantifying how territorial frogs return after experimental mainly consists of primary lowland rainforest bordering removals. It is also one of few studies addressing amphib- the ‘Arataye’ river to the south. ian orientation by quantifying full homing trajectories. Even though the HDF method has its clear limitations Telemetry experiment such as short detection range (10–30 m) and no possibility We used the harmonic direction-finding (HDF hereafter) to identify individuals based on the signal, we demonstrate telemetry technique to obtain homing trajectories of ex- that it can be a useful tool for understanding movement perimentally translocated territorial males. The HDF sys- patterns of small animals in complex environments, such tem consists of a directional transceiver and a passive as the tropical rainforest. Several authors have reported reflector. The transceiver emits and recaptures a radio problems, such as altered behavior and injuries, from ex- signal which gets reflected from the tag attached to an ternal attachment methods on anurans [28,31,50]. Our at- animal, thereby providing directional information (for tachment method was generally suitable for a short-term more details see [26,29]). The miniature size of the re- (up to five days) tracking but could be problematic for flector tags allows this technique to be used on smaller longer-term studies. Frogs exhibited all their natural re- animals than permitted by conventional active radio productive behaviors such as calling, territorial defense, tracking. We used a commercially available transceiver courtship, and on one occasion even tadpole transport (R8, RECCO® Rescue System, Lidingö, Sweden) with re- while having a reflector attached. However, on some occa- flector tags consisting of a Schottky diode soldered be- sions, we observed skin abrasions around the waist after tween two antennas. Antennas were made of 40 μm removing the tag. Observed injuries healed in a few days steel strands forming a 2 cm by 10 cm T-shaped dipole and at least in some cases we know that males continued with the braze point sealed in a heat-shrink tubing. We to defend territories for weeks and months following the attached the reflectors to the frogs using a waistband tracking. made of 1 mm diameter silicon tubing similarly to Gourret and Schwarzkopf [27]. The short part of the T- Conclusions shaped antenna was secured inside of the tube and the Overall, our results underline the importance of quanti- waistband was fixed with a cotton thread going through fying the individual trajectories after experimental dis- the inside of the tubing. The tag together with the at- placements in understanding amphibian orientation. tachment made up for less than 5% of frogs’ body weight Taken together with measures of initial homeward orien- (rangefrog weight = 1.4 – 2.3 g, maxtag weight = 0.07 g). tation, they reveal a very accurate homing behavior in A. The telemetry experiment took place from 21 January femoralis. The straightness of trajectories and initial 2013 to 15 March 2013. During this period, 18 territorial homeward orientation suggest an integration of learned males were captured and equipped with reflector tags. Pašukonis et al. Frontiers in Zoology 2014, 11:29 Page 6 of 9 http://www.frontiersinzoology.com/content/11/1/29

Calling males were detected and identified as territorial them. Because the harmonic signal does not carry an in- if they showed stereotypical territorial defense behavior dividual signature and handling would be necessary for (calling and phonotactic approach), which was elicited identification, we never translocated more than one frog by broadcasting conspecific advertisement calls, simulat- at a time in the same area. ing an intruder. Frogs were captured with transparent airtight plastic bags, photographed for identification, and Trajectory analysis their precise capture positions were recorded with the Initial visualization, extraction of coordinates and distance mobile GIS software ArcPad™ 10.0 (ESRI, Redlands, CA, measurements were done in the GIS software ArcGIS™ 10 USA) on pocket computers (MobileMapper™ 10, Spectra- (ESRI, Redlands, CA, USA). The geographic coordinates Precision, Westminster, CO, USA) using a detailed back- of all locations were projected (UTM, zone 22 N, WGS84) ground map based on a grid of reference points and and extracted as X- and Y-coordinates in meters. We natural structures. Each frog was equipped with a reflector grouped points spaced less than two meters apart to a sin- and immediately released at its initial capture position, gle position, as they might have fallen within the measure- where it was observed for a 24 h period to confirm normal ment error of the exact position in the field. territorial behavior with the reflector. Each frog was lo- We calculated a straightness coefficient (SC) for the cated at least once during this period and confirmed to path between the release site and the home territory as behave territorially if it showed any of the following be- the ratio between the straight-line distance and the actual haviors: calling, phonotactic approach of a simulated in- path distance, with a ratio of one indicating a perfectly truder, courtship. Two males were excluded at this stage straight path. To test for initial homeward orientation, we and their tags were removed because no territorial behav- considered the bearing between the release point and the ior was observed, and another individual lost the waist- first position further than two meters away from the band and could not be located again. release point (n = 8, meandistance to first position =6.6m, When normal territorial behavior with the tag was SD = 4.2). Significant homeward orientation was tested confirmed, each frog was captured with an airtight plas- using Rayleigh’s test for unimodal distribution. Two indi- tic bag, placed in an opaque container and translocated viduals were excluded from this analysis because the first 50 m away from the home territory (n = 15). Each animal recovery positions were too far (> 20 m) from the release was designated to one of five displacement directions point to be seen representative for initial orientation. The (N, E, S, W and NW), avoiding terrain and vegetation Rayleigh test was performed with a circular statistics pro- where tracking might have been impossible. The release gram Oriana 4.02 (Kovach Computing Services, Pentraeth, points were located and marked using the same detailed Wales, UK). background map on a pocket computer. Allobates femoralis, like most dendrobatid frogs, is diurnal and does not Return time analysis move during the dark hours, which was confirmed by To analyze return times, we considered diurnal activity preliminary observations and during this study. There- hours from 7 00 h to 19 00 h. We further split the total fore, individuals were tracked during the daylight hours return time of each individual into time actively moving (07 00 h to 19 00 h) until they returned to their territor- and time stationary. Any time interval where an individual ies. Locations were visited approximately every hour moved further than two meters between two consecutive with occasional shorter intervals when the frog was relocalizations was considered as time active. We com- found actively moving. Longer intervals were sometimes pared the two activity modes using the Wilcoxon Signed- forced by bad weather conditions and/or additional time Rank test for comparison of paired samples using the R taken to relocate the frog. Stats package 3.0.1 [51]. To locate the frogs, we followed the increasing ampli- tude of the reflected signal until visually spotting an in- Arena assay dividual. In cases of poor visibility or if an individual We further assessed the initial orientation ability of male was hiding in the leaf litter, we narrowed the signal to A. femoralis in an experimental arena setup. A circular less than one meter. If an individual remained stationary arena with a diameter of 240 cm and 40 cm height was and hidden for longer periods, we carefully uncovered set up at an arbitrarily chosen location in the natural habi- the frog at least once a day to make sure that the tag tat. The walls of the arena where made out of blue plastic had not fallen off and that the individual had no injuries. tarpaulin and supported by fixed poles, while the floor Occasionally, the tag had twisted to the side or under- was covered with an off-white, thick, plastic sheet, marked neath the frog, in which case the frog was carefully ma- with a grid as a reference for position measurements. nipulated by pulling the antenna to reposition the tag. To test for initial orientation after experimental trans- These manipulations never took more than a minute location, territorial males were caught in their territories and frogs never moved more than a meter as a result of and released in the arena, following the same catching Pašukonis et al. Frontiers in Zoology 2014, 11:29 Page 7 of 9 http://www.frontiersinzoology.com/content/11/1/29

and translocation protocol as in the telemetry trials. Sin- gle individuals were placed under a release device in the center of the arena and left to habituate for 15 min. The release device consisted of two upturned plastic flowerpots, stacked one over the other. After the outer part was lifted with a string, the frog could leave the inner part to all directions through cutout exit holes. This release design was necessary because some individuals showed immediate escape behavior when left exposed with no cover. Trials were filmed for 90 min without ex- perimenter presence using a wide angle video camera (GoPro™ HD HERO2, Woodman Labs Inc., San Mateo, CA, USA) on a tripod. After leaving the release device and reaching the edge of the arena, frogs usually circled and eventually climbed over the wall. Individuals were left to home back to their territories. Frogs that did not exit the release device within 90 min were returned to their home territories and excluded from the analysis. Fifty-six arena trials were conducted from 19 January 2013 to 30 March 2013. Nine trials were excluded for Figure 4 Arena trajectory. Example of a single coded trajectory technical problems (n = 3) or because frogs did not leave from the arena experiment, using a graphical arena representation the release device within 90 min (n = 6). Comparisons of as an interface. Solid outer circle represents the arena wall, small individual photographs revealed that nine individuals filled gray circle represents the release device, black dots connected where tested twice and in such cases only the first trial by an interpolation line represent each hop of the frog in the arena. The orientation bearing was measured at the outer dashed circle was included in the analysis. (100 cm) and the straightness coefficient (SC) was calculated for the Initially, we translocated frogs that had their territories path between inner (30 cm) and outer (100 cm) dashed circles. within 30–70 m around the arena to enable a direct comparison with the homing trajectories of the frogs that were used in the tracking experiment where they SC was calculated as the ratio between the straight line were translocated over distances of 50 m. Subsequently, and the actual path distance in the arena. The expected we included frogs closer (5–30 m) and further away homeward orientation bearings were calculated from the (70–105 m) from the arena to assess the effects of trans- coordinates of each capture position as recoded on the location distance on the accuracy of initial orientation. digital map and the center of the arena. To test for significant homeward orientation in different Initial orientation analysis distance groups, we used the second order Hotelling’scir- In total, we analyzed 38 valid arena trials. We used a cular test for significant unimodal distribution of each custom script written by the first author for MATLAB sample. In addition, we looked for significant differences 7.11.0.584 (The MathWorks Inc., Natick, MA, USA) to between distributions obtained from displacements over code and extract the trajectory of each frog in the arena. different distances, by using the Hotelling’stwosample A graphical representation of the arena with the refer- test. For significant mean directions a 95% confidence ence grid was simultaneously displayed with the video interval was calculated. Statistical analyses were performed recording of each trial and each hop position of the frog with Oriana 4.02. was coded with a mouse click at the corresponding loca- Our study was approved by the scientific committee of tion in the graphical representation. Each position was the research station where fieldwork was conducted. All exported as the corresponding X- and Y-coordinates necessary permissions were provided by the ‘Centre Na- with the origin at the center of the arena. Arena trajec- tional de la Recherche Scientifique’ (CNRS) and by the tories were analyzed between an inner circle with 30 cm ‘Direction Régionale de l’Environment de Guyane’ radius and an outer circle with 100 cm radius, in order (DIREN). All sampling was conducted in strict accord- to avoid the wall influenced movements close to the re- ance with current French and EU law and followed the lease device and close to the arena wall (Figure 4). ASAB guidelines for the treatment of animals in behav- We used the coordinates to calculate the orientation ioral research and teaching. bearing of each individual at 100 cm away from the center and the SC of their trajectory from 30 cm to 100 cm Competing interests away from the center. Similar to the telemetry trials, the The authors declare that they have no competing interests. Pašukonis et al. Frontiers in Zoology 2014, 11:29 Page 8 of 9 http://www.frontiersinzoology.com/content/11/1/29

Authors’ contributions 21. Tracy CR, Dole JW: Orientation of displaced California Toads, Bufo boreas, AP designed the study, collected, analyzed, and interpreted the data and to their breeding sites. Copeia 1969, 4:693–700. drafted the manuscript. MCL participated in the design of the study, data 22. Grubb JC: Orientation in post-reproductive Mexican Toads, Bufo valliceps. collection, analysis, and interpretation. LL participated in the statistical analysis Copeia 1970, 4:674–680. and interpretation of the data. MR participated in the design of the study and 23. van Nuland GJ, Claus PFH: The development of a radio tracking system advised on the GIS data processing. WH participated in the design of the study. for anuran species. Amphibia Reptilia 1981, 2:107–116. MCL, LL, MR and WH reviewed and provided valuable comments on the 24. Stouffer RHJ, Gates JE, Hocutt CH, Stauffer JRJ: Surgical implantation of a manuscript. All authors read and approved the final manuscript. transmitter package for radio-tracking endangered hellbenders. Wildlife Soc B 1983, 11:384–386. Acknowledgments 25. Riley JR, Smith AD, Reynolds DR, Edwards AS, Osborne JL, Williams IH, We are grateful to Prof. Dr. Nicolas Perrin for providing the telemetry Carreck NL, Poppy GM: Tracking bees with harmonic radar. Nature 1996, equipment and to the staff of CNRS Guyane for the logistic support during 379:29–30. the fieldwork. We also thank Dr. Daniel Bowling for helping with writing the 26. Mascanzoni D, Wallin H: The harmonic radar: a new method of tracing MATLAB script and Prof. Dr. Ulrich Sinsch along with two anonymous reviewers insects in the field. Ecol Entomol 1986, 11:387–390. for their valuable comments and suggestions to the manuscript. This work was 27. Gourret A, Alford R, Schwarzkopf L: Very small, light dipole harmonic tags supported by the Austrian Science Fund (FWF): projects W1234-G17 and for tracking small animals. Herpetol R 2011, 42:522–525. P24788-B22 (PI: Eva Ringler). 28. Langkilde T, Alford RA: The tail wags the frog: harmonic radar transponders affect movement behavior in Litoria lesueuri. J Herpetol Author details 2002, 36:711–715. 1Department of Cognitive Biology, University of Vienna, Althanstrasse 14, 29. Rowley JJL, Alford RA: Techniques for tracking amphibians: the effects of 1090 Vienna, Austria. 2Department of Biological Sciences, Virginia Tech, tag attachment, and harmonic direction finding versus radio telemetry. Blacksburg, VA 24061, USA. 3Department of Integrative Zoology, University of Amphibia Reptilia 2007, 28:367–376. Vienna, Althanstrasse 14, 1090 Vienna, Austria. 4Department of Tropical 30. Leskovar C, Sinsch U: Harmonic direction finding: a novel tool to monitor Ecology and Animal Biodiversity, University of Vienna, Rennweg 14, 1030 the dispersal of small-sized anurans. Herpetol J 2005, 15:173–180. Vienna, Austria. 31. Pellet J, Rechsteiner L, Skrivervik AK, Zurcher JF, Perrin N: Use of the harmonic direction finder to study the terrestrial habitats of the Received: 31 December 2013 Accepted: 19 March 2014 European tree frog (Hyla arborea). Amphibia Reptilia 2006, 27:138–142. Published: 25 March 2014 32. Weygoldt P: Evolution of parental care in dart poison frogs (Amphibia: Anura: Dendrobatidae). J Zool Syst Evol Res 1987, 25:51–67. References 33. Brown JL, Morales V, Summers K: Home range size and location in relation 1. Walcott C: Pigeon homing: observations, experiments and confusions. to reproductive resources in poison frogs (Dendrobatidae): a Monte – J Exp Biol 1996, 199:21–27. Carlo approach using GIS data. Anim Behav 2009, 77:547 554. 2. Müller M, Wehner R: Path integration in desert ants, Cataglyphis fortis. 34. Summers K, Sea McKeon C, Heying H: The evolution of parental care and Proc Natl Acad Sci U S A 1988, 85:5287–5290. egg size: a comparative analysis in frogs. Proc R Soc Lond B Biol Sci 2006, – 3. Boulenger GA: Some remarks on the habits of British Frogs and Toads, 273:687 692. with reference to Mr. Cummings's recent communication on Distant 35. Amézquita A, Lima AP, Jehle R, Castellanos L, Ramos O, Crawford AJ, Gasser Orientation in Amphibia. P Zool Soc Lon 1912, 82:19–22. H, Hodl W: Calls, colours, shape, and genes: a multi-trait approach to the 4. Cummings BF: Distant orientation in Amphibia. PZoolSocLon1912, 82:8–19. study of geographic variation in the Amazonian frog Allobates femoralis. – 5. Ferguson DE, Landreth HF: Celestial orientation of Fowler's toad Bufo Biol J Linn Soc 2009, 98:826 838. fowleri. Behaviour 1966, 30:105–123. 36. Roithmair ME: Territoriality and male mating success in the dart-poison frog, – 6. Sinsch U: Orientation and navigation in Amphibia. Mar Freshw Behav Phy Epipedobates femoralis (Dendrobatidae, Anura). Ethology 1992, 92:331 343. 2006, 39:65–71. 37. Ringler M, Ursprung E, Hodl W: Site fidelity and patterns of short- and 7. Bogert CM: Results of the Archbold Expeditions No. 57. A field study of long-term movement in the brilliant-thighed poison frog Allobates – homing in the Carolina toad. Am Mus Novit 1947, 1355:1–24. femoralis (Aromobatidae). Behav Ecol Sociobiol 2009, 63:1281 1293. 8. Dole JW: Homing in Leopard Frogs Rana pipiens. Ecology 1968, 49:386–399. 38. Ringler E, Pašukonis A, Hodl W, Ringler M: Tadpole transport logistics in a 9. Jameson DL: Population structure and homing responses in the Pacific Neotropical poison frog: indications for strategic planning and adaptive tree frog. Copeia 1957, 3:221–228. plasticity in anuran parental care. Front Zool 2013, 10:1–9. 10. Pašukonis A, Ringler M, Brandl HB, Mangione R, Ringler E, Hodl W: The 39. Menzel R, Greggers U: Guidance by odors in honeybee navigation. homing frog: high homing performance in a territorial dendrobatid frog J Comp Physiol A 2013, 199:867–873. Allobates femoralis (Dendrobatidae). Ethology 2013, 119:762–768. 40. Matthews KR, Pope KL: A telemetric study of the movement patterns and 11. Twitty V, Grant D, Anderson O: Long distance homing in the newt Taricha habitat use of Rana muscosa, the mountain yellow-legged frog, in a rivularis. Proc Natl Acad Sci U S A 1964, 51:51–58. high-elevation basin in Kings Canyon National Park, California. J Herpetol 12. Grubb JC: Olfactory orientation in southern leopard frogs, Rana 1999, 33:615–624. utricularia. Herpetologica 1975, 31:219–221. 41. Able KP: The debate over olfactory navigation by homing pigeons. JExpBiol 13. Ishii S, Kubokawa K, Kikuchi M, Nishio H: Orientation of the toad, Bufo 1995, 199:121–124. japonicus, toward the breeding pond. Zool Sci 1995, 12:475–484. 42. Lohmann KJ, Lohmann CMF, Putman NF: Magnetic maps in animals: 14. Oldham RS: Orienting mechanisms of green frog Rana clamitans. Ecology nature's GPS. J Exp Biol 2007, 210:3697–3705. 1967, 48:477–491. 43. Phillips JB: Magnetic navigation. J Theor Biol 1996, 180:309–319. 15. Diego-Rasilla FJ, Luengo RM, Phillips JB: Magnetic compass mediates 44. Hodl W: Dendrobates femoralis (Dendrobatidae): a handy fellow for frog nocturnal homing by the alpine newt, Triturus alpestris. Behav Ecol bioacoustics.InProceedings of the 4th Ordinary General Meeting of the Sociobiol 2005, 58:361–365. Societas Europaea Herpetologica: 17–21 August 1987; Nijmegen. Edited by van 16. Fischer JH, Freake MJ, Borland SC, Phillips JB: Evidence for the use of Gelder JJ, Strijbosch H, Bergers PJM. 1987:201–204. magnetic map information by an amphibian. Anim Behav 2001, 62:1–10. 45. Ursprung E, Ringler M, Hodl W: Phonotactic approach pattern in the 17. Landreth HF, Ferguson DE: Newt orientation by sun-compass. Nature 1967, neotropical frog Allobates femoralis: a spatial and temporal analysis. 215:516–518. Behaviour 2009, 146:153–170. 18. Ferguson DE: The sensory basis of orientation in amphibians. Ann N Y Acad Sci 46. Hodl W, Amézquita A, Narins PM: The rôle of call frequency and the auditory 1971, 188:30–36. papillae in phonotactic behavior in male dart-poison frogs Epipedobates 19. Matthews KR: Response of mountain yellow-legged frogs, Rana muscosa, femoralis (Dendrobatidae). JCompPhysiolA2004, 190:823–829. to short distance translocation. J Herpetol 2003, 37:621–626. 47. Ursprung E, Ringler M, Jehle R, Hodl W: Strong male/male competition 20. Sinsch U: Orientation behaviour of toads (Bufo bufo) displaced from the allows for nonchoosy females: high levels of polygynandry in a territorial breeding site. J Comp Physiol A 1987, 161:715–727. frog with paternal care. Mol Ecol 2011, 20:1759–1771. Pašukonis et al. Frontiers in Zoology 2014, 11:29 Page 9 of 9 http://www.frontiersinzoology.com/content/11/1/29

48. Bee MA: Selective phonotaxis by male wood frogs (Rana sylvatica) to the sound of a chorus. Behav Ecol Sociobiol 2007, 61:955–966. 49. Ursprung E, Ringler M, Jehle R, Hodl W: Toe regeneration in the neotropical frog Allobates femoralis. Herpetol J 2011, 21:83–86. 50. Bartelt PE, Peterson CR: A description and evaluation of a plastic belt for attaching radio transmitters to Western Toads (Bufo boreas). Northwest Nat 2000, 81:122–128. 51. R Development Core Team: R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013. http://www.R-project.org/.

doi:10.1186/1742-9994-11-29 Cite this article as: Pašukonis et al.: Homing trajectories and initial orientation in a Neotropical territorial frog, Allobates femoralis (Dendrobatidae). Frontiers in Zoology 2014 11:29.

Submit your next manuscript to BioMed Central and take full advantage of:

• Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit